A predictive signal is not the same thing as a profitable trade. Between research and realised PnL sit liquidity, timing, transaction costs, market impact and the quality of execution.
Quantitative research often receives most of the attention in discussions about systematic investing. That makes sense, because without a useful prediction there may be nothing worth trading.
But a research result only becomes commercially meaningful once the strategy interacts with a live market, and that transition introduces a different set of problems. Execution can preserve an edge. It can improve how efficiently the edge is captured. Or it can remove much of the value that appeared to exist in research.
A forecast is not a fill
A backtest can assume a theoretical transaction at a particular price. A live trading system has to obtain that price from another market participant, and that difference creates friction.
Bid-offer spreads, fees, slippage, latency, market impact, partial fills, changing liquidity and adverse selection can all look small individually. Across a high-turnover strategy they can materially change the economics.
The relevant question is therefore not only whether a signal predicts returns. It is whether those returns remain after the strategy has actually been traded.
Execution is part of monetisation
This is why traders inside quantitative organisations can be much more than order handlers. They may think about how capital is deployed across signals, how orders interact with available liquidity, when to trade aggressively and when to wait, and how execution algorithms behave in different regimes.
They also watch where transaction costs are increasing and whether market impact is changing as the strategy scales. Those decisions help determine how much theoretical alpha becomes realised PnL.
In highly systematic environments, much of the process may be automated. That does not make execution irrelevant. It changes the form of the execution problem, moving the work into algorithm design, portfolio optimisation, market microstructure, monitoring and continuous analysis of realised trading outcomes.
Live trading can improve the research
Execution feedback can also change the research process itself. Suppose a signal performs well before costs but poorly after implementation. That may indicate the opportunity is too short-lived, that the portfolio is trading too much, or that the universe is too illiquid.
The signal may be crowded, the execution approach may need improvement, or the original research assumptions may simply have been unrealistic. Those findings should not remain isolated inside a trading team. They are information about the strategy, and strong research environments allow that feedback to flow back into model development.
This creates crossover roles
It also helps explain why job titles around quantitative trading can become difficult to define. A Quantitative Trader may perform research. A researcher may work deeply on transaction-cost modelling. A Quant Developer may build the simulation that determines whether an idea is viable, and an engineer may improve a trading path enough to change the economics of a strategy.
Each person is working on a different part of the same conversion process: prediction into position, position into execution, execution into realised outcome. That is why ownership often tells you more than title.
What this means in hiring
When hiring somebody close to production trading, it is useful to understand how they think about this conversion. For researchers: how were transaction costs modelled, how close were research assumptions to live execution, and what changed after deployment?
For traders: how did they measure execution quality, what did they change when costs increased, and how closely did they work with research and engineering? For engineers: which systems did they own, what performance constraint applied, and did an infrastructure improvement change trading outcomes? These are different questions, but they all connect technical work to commercial impact.
The best research still needs a path to market
None of this diminishes the importance of alpha research. Execution cannot manufacture a durable prediction from nothing. But the reverse is also true: a strong prediction that cannot be implemented economically may never become a useful strategy.
Systematic investing therefore works best when research, portfolio construction, trading and engineering are treated as connected problems. The hand-off between them is where a surprising amount of value can be created or lost.
Key takeaway
The commercial value of quantitative research depends partly on how effectively it can be expressed in live markets. Execution quality, transaction costs, liquidity and market impact are not downstream details. They are part of the investment process.